Why is interest expense excluded from unlevered free cash flow, and where in the DCF is the cost of debt (including its tax benefit) actually captured?

What the interviewer is actually testing

UFCF questions are where interviewers separate candidates who memorized a formula from candidates who understand what cash flow is. Three layers get tested.

Layer 1: the formula, stated with reasons. "How do you calculate unlevered free cash flow?" Anyone can recite EBIT(1−t) + D&A − CapEx − ΔNWC. The strong candidate attaches a one-clause rationale to each term without being asked: EBIT because it's pre-financing, taxed to get NOPAT, D&A back because non-cash, CapEx out because it's the real reinvestment, change in NWC because growth ties up cash. That version takes fifteen extra seconds and completely changes the impression.

Layer 2: the boundary tests. Expect: "Why don't you subtract interest?" (financing item; the tax shield lives in WACC), "Why tax EBIT when the company pays less tax than that?" (avoiding a double count of the interest tax shield), "Is depreciation really irrelevant to value?" (no: it shields taxes; a company with more D&A has higher UFCF, all else equal), and "Walk me from net income to UFCF" (add back D&A and after-tax interest, subtract CapEx and ΔNWC). Each one probes whether you know where the line sits between operating and financing.

Layer 3: forecast judgment. Elite boutiques push into "how would you actually project this?": what drives revenue, why margins fade or expand, what the terminal year must look like. They're checking whether you understand that the terminal year has to be steady-state (growth near GDP, CapEx converging to just above D&A) because the terminal value capitalizes whatever you leave there.

Signal mastery by volunteering the consistency checks: "I'd make sure reinvestment scales with growth: you can't grow revenue 10% on maintenance CapEx," and "by the final year growth fades to 2–3% and CapEx sits just above D&A." Candidates who talk about their forecast as an internally consistent economic story, rather than a stack of independent percentages, sound like they've built and defended real models.

Common traps

Trap 1: Subtracting interest expense. The most common error in the entire DCF chapter. Interest is a financing flow; including it makes the cash flow levered, breaks the pairing with WACC, and double-counts the cost of debt.

Say it out loud: "I exclude interest entirely: UFCF is capital-structure-neutral. The cost and tax benefit of debt are captured in WACC through the after-tax cost of debt, not in the cash flows."

Trap 2: Using the level of NWC instead of the change. Candidates subtract total net working capital rather than the year-over-year increase. Only the incremental cash tied up in the operating cycle affects that year's cash flow.

Say it out loud: "I subtract the increase in net working capital: if NWC goes from $100M to $110M, that's a $10M use of cash; if it declines, the release of working capital is a source of cash."

Trap 3: Forgetting that D&A's add-back is not a free pass. Some candidates conclude D&A "doesn't matter" since it's added back. It matters through taxes: D&A reduces EBIT, which reduces taxes, and only then gets added back, so higher D&A means higher UFCF via the tax shield.

Say it out loud: "D&A is non-cash, but it isn't value-neutral: every extra dollar of D&A saves the tax rate times a dollar in cash taxes. That's why we deduct it before taxing EBIT and add it back after."

Trap 4: Botching the net-income bridge. Asked to go from net income to UFCF, candidates add back D&A and stop, or add back gross interest. Net income already deducted interest and got a tax break on it, so you must add back interest × (1 − t).

Say it out loud: "From net income: add back D&A, add back after-tax interest expense (interest times one minus the tax rate, since unlevering also removes the interest tax shield), then subtract CapEx and the increase in NWC."

Trap 5: Taxing EBIT at the company's low effective rate forever. Using a temporarily depressed effective tax rate (credits, NOLs, one-time items) in the terminal year capitalizes a temporary benefit into perpetuity.

Say it out loud: "For near years I'd use the effective rate the company actually expects to pay; by the terminal year I revert to the marginal statutory rate, because temporary tax items shouldn't be capitalized into perpetuity."

Trap 6: A terminal year that isn't steady-state. Leaving 12% growth, expanding margins, or CapEx far above D&A in the final explicit year, then applying a perpetuity formula to it. The terminal value freezes whatever economics the final year shows.

Say it out loud: "Before applying a terminal value, I make sure the final year is steady-state: growth faded to roughly 2–3%, stable margins, and CapEx converged to just above D&A. Otherwise the perpetuity capitalizes unsustainable economics."

Trap 7: Ignoring stock-based compensation. Treating SBC like D&A (a pure non-cash add-back) quietly overstates UFCF, because SBC is real economic compensation that dilutes shareholders.

Say it out loud: "SBC is non-cash but not costless: the clean treatments are either to leave it as an expense and not add it back, or to add it back and then account for the dilution in the share count. Adding it back and ignoring dilution overstates value."